Hospitals — how this layer is made

What you are looking at

Each point is a hospital — a place with beds, as tagged in OpenStreetMap. 210,666 of them worldwide, coloured by whether they have an accident and emergency department and sized by the area of the site.

Click any hospital to see what OpenStreetMap records about it — who runs it, where a mapper noted that. Where it has a Wikipedia article, a short summary is fetched live and shown in your own language; and every hospital links to the OpenStreetMap object it came from.

The layer that reverses this site's usual warning

Read this before the map, because on every other layer here the equivalent paragraph says the opposite.

Counted inside each country's boundary, per million people: Japan 83.61, Australia 45.71, France 38.99, India 38.57, Brazil 37.15, Germany 27.63, the United Kingdom 26.85, the United States 23.22, Nigeria 18.80, Indonesia 14.80, China 9.05.

India is ahead of Germany, and Nigeria is ahead of China. Germany against India works out at 0.72, where the site's cultural layers run past 100 the other way. The densest single view on this map anywhere on earth is in Tamil Nadu, and the second is Seoul.

HueMaps' layers usually warn that an empty region may mean nobody has mapped it rather than that nothing is there. That warning is real, and this layer is the reason it is credible rather than reflexive: where a subject genuinely is mapped everywhere, this site's maps show it. A hospital is a landmark in every country — signposted, named, known to everyone nearby — so it gets mapped in every country.

None of which makes the rate a measure of healthcare. A country with many small district hospitals will score above one with a few large ones, and this map counts sites rather than beds, staff or care.

Mapped, not modelled

Each hospital is an object a mapper drew, from survey, local knowledge or aerial imagery. Nothing here is modelled or estimated, and no national health register is involved.

What counts, and what does not

Two tags: `amenity=hospital` (216,942 objects) and `healthcare=hospital` (a further 3,847 that carry it without the first). The second selection is worth having for those alone — `healthcare=*` is the newer scheme, the two are used side by side, and taking only the older key would lose a set that grows every year.

Not clinics, and not doctors' surgeries. `amenity=clinic` and `amenity=doctors` are far more numerous and are a different institution: somewhere you are seen and sent home. This map draws the places with beds, because that is the question a reader is asking.

Where a hospital is mapped both as a point and as the outline of its plot — which is common here — the two are merged into one feature rather than drawn twice.

The three colours, and why the grey one is the biggest

Each hospital is coloured by whether OpenStreetMap says it has an accident and emergency department — the tag is `emergency`, and it takes `yes` or `no`.

Emergency department — 24,878 hospitals.
No emergency department — 6,434, where somebody has explicitly recorded that it has none.
Not recorded — everything else, and that is 85% of the layer.

Read the grey carefully, because it is a statement about the record and not about the hospital. A grey dot is not a hospital without a casualty department. It is a hospital where nobody has said either way, and there are far more of those than of the other two put together.

Where the grey is, and why that matters more here than anywhere

The share of hospitals carrying the tag at all, by region:

| region | tagged | |---|---| | North America | 71.4% | | Oceania | 42.8% | | Western Europe | 34.3% | | Latin America | 25.5% | | South-East Asia | 15.0% | | Sub-Saharan Africa | 7.8% | | East Asia | 5.5% | | South Asia | 3.7% |

North America is nineteen times South Asia — and South Asia holds 62,094 of this map's hospitals, more than any other region.

That collides with the best thing about this layer. Everywhere else on this page you are told that hospitals are the one subject mapped evenly across the world, with India ahead of Germany per head. The emergency tag is not: it is a North American and European habit that has not spread. So the colours describe the world's hospitals honestly only where the colour exists.

If you untick the grey row in the legend, what is left is a map of North America, Europe and Australia. That is not where the emergency departments are. It is where they have been written down, and it is the plainest demonstration of the caveat this page could give you.

What is deliberately not coloured

The division you would probably want first is public against private, and OpenStreetMap has a tag for it: `operator:type`. It is on a small enough minority that colouring by it would draw a map of where that tag has been used rather than of who runs the world's hospitals, with nothing to tell you which you were looking at. `healthcare:speciality` is thinner still and often lists several specialities at once.

The emergency tag is admitted where those are refused for one reason: the grey class shows you the gap. Two confident colours over a third of the data would hide it.

Size

Each symbol is sized by the area of the hospital site, in square metres measured on the globe. Symbol area is proportional to that (radius to its square root), so a site four times the area draws twice as wide.

Here that means something. A hospital's plot is a fair proxy for its size in a way a library's is not: wards, theatres, car parks and access roads all scale roughly with the place. 49.9% of hospitals are mapped as an outline, and among those the median covers 6,977 m². The other half are recorded as a single point with no shape, and are drawn at a nominal 7,000 m² rather than at the minimum — being mapped as a point says something about the mapper's time, not about the size of the hospital.

92.6% carry a name.

The largest dots are psychiatric institutions, not major general hospitals. The top of the range is Napa State Hospital in California at nearly 6 km², then Hospital Interzonal Doctor Domingo Cabred in Argentina, Colonia Doctor Bernardo Etchepare in Uruguay and Sterkfontein in South Africa. These are colonies built on large grounds, and the grounds are counted. The biggest dots on this map are the ones with the most land rather than the most beds.

A picture of the present

Today's map, not a history. OpenStreetMap is edited continuously; this copy is refreshed periodically. There is no year control, and a hospital closed some time ago may still be mapped until a mapper updates it.

Source and licence

Back to the map